The peak of the outbreak would come in the middle of April, task force member Deborah Birx explained. She showed a graph using modeling from the University of Washington’s Institute for Health Metrics and Evaluation (IHME). The projected peak would come on April 15, with 2,214 deaths on that day.
Late Wednesday night, the IHME released revised estimates, based on new data. During the first wave of the epidemic, its model projects, the death toll will be 93,765 — an increase of 14 percent from its model the previous day. That’s just the first wave, looking at the number of deaths through July. In the fall and winter, the virus is expected to reemerge and pose a significant threat once again.
That shift is a function in part of things looking more bleak over the short term. The new model suggests that the number of deaths each day is now most likely to peak on April 16, a day later than suggested Monday, with 400 more deaths that day.
In other words, the data from Monday have been shifted upward significantly — particularly the worst-case scenario represented by the upper-bound of the uncertainty area displayed on the graph. The graph presented from the White House on Monday stopped at about 3,500 deaths a day. The new estimate projects a possible worst-case peak of 4,400 deaths on April 21.
Why the shift? Models of pandemics like the coronavirus outbreak, as FiveThirtyEight explained on Tuesday, are very tricky. What’s more, new data mean a shift in what’s projected. Just as models of the likely winner of the Democratic primary shifted dramatically after former vice president Joe Biden won South Carolina, new data fed into coronavirus models can change the expected outcome.
“What we do is that every time we get more data, you feed it back in and relook at the model. Is the model really telling you what is actually going on?” task force member Anthony S. Fauci said Monday. “Models are as good as the assumptions you put into them, and as we get more data, then you put it in and that might change,” he added.
One bit of data added to the IHME models was the actual number of deaths over the past few days. The graph presented on Monday, for example, included an estimate that there would be about 850 deaths on April 1. Data from Johns Hopkins University indicates that the number was closer to 890 — though even that figure includes some uncertainty. The model released on Wednesday night estimated there would be 899 deaths on April 1.
President Trump had speculated that life would be returning to some semblance of normal by June 1. By then, the IHME data suggest, about 91,000 people will have died of covid-19, the disease caused by the coronavirus. If the model continues to be revised upward, the figure could approach 170,000 — just from the first three months of deaths.
An important factor in the IHME data is how the peaks in deaths will be staggered around the country. New York, currently the main flash point of the outbreak in the United States, is expected to peak at 855 deaths on April 10. Alabama is projected to peak a bit later, with 303 deaths on April 19. (Its governor was right: Alabama is no California, where the model projects a peak of 119 deaths on April 28.)
Those are median estimates. On the upper bound of possible deaths, late April and early May would see a series of staggered peaks from big states.
Those, again, are worst-case scenarios. But they, too, will be revised as new data come in, and those revisions could be upward.
In an interview with the New York Times, Fauci offered one bit of good news, in a way.
“You will continue to see deaths at a time when you have actually very good control of the new infections and the outbreak itself,” he said.
Some consolation for those who have survived, anyway.